Agentic artificial intelligence (AI) has the potential to boost the workforce for travel companies, but the importance of getting the basics right is not to be underestimated.
Fundamentals such as security, data and financial control need to be properly thought out as companies scale the technology, according to Trainline CTO Mike Hyde.
At the recent TravelTech Show, Hyde discussed the platform's AI journey alongside Harry Herbert, enterprise go-to-market lead for Anthropic.
Hyde suggested imagining a world where Trainline had 3,000 engineers versus the roughly 600 it currently has, with that workforce boost mostly agentic.
"The difference is we suddenly have 3,000 engineers, and all of our systems are designed to support 600 engineers, and the 3,000 engineers are less trustworthy than the 600," Hyde said. "You can't assume they're going to do the right thing all day, every day. Then, what kind of a security system would you need to run a 3,000+ company where you don't trust all the people?"
He added that this is the "mental model" Trainline is trying to figure out and described it as a good place to start for many businesses.
"What would you do if you were 5x as big as you are, with a whole series of very low-trust employees?"
Focusing on the fundamentals, including process, design, governance, security and financial control, also helps companies navigate the rapidly evolving landscape, according to Hyde.
As part of the discussion, he touched on the cost of AI and tokens and said that although it's early days, the industry needs to figure out the cost of compute going forward.
"We're starting to use a system of individual budgets for engineers, and then maybe it's special budgets for certain projects we know will be expensive token-wise. But it might be worth it because we're building something new."
Hyde said all these fundamentals might not be seen as sexy or headline-grabbing but added: "They're the foundation that lets you adopt whatever the latest model is. The fundamentals don't change as fast as the models do."
Herbert also stressed the importance of putting processes and evaluation practices in place "to judge whether a new feature is actually relevant to the use cases and customers you're serving."
AI in action
Trainline has been building AI models on top of its data platform and seen the technology's impact in three different areas—building more intelligent customer products using its data, understanding change in the way people search and discover travel and adapting to it and how AI changes internal operations.
He added that the company has been implementing internal AI capabilities for about 18 months but has seen an "inflection point" in the past three to four months. Hyde said workloads have shifted, enabling the company to turn to more ambitious internal initiatives.
"We've come to the point of view that it's not enough to just roll AI out; you get to a point where you step back and realize we need to rethink things more fundamentally, like how will we work in the future," he said.
Trainline's initial approach to roll out tools and see what happens is likely where many companies start on their AI journey, Hyde said, adding that it resulted in 10% to 20% productivity gains.
He said the company is now taking a more top-down view to build automations that don't just save someone an hour or two but take on an entire task or job, using a series of agentic steps.
Herbert said he sees a mix among travel companies introducing AI from the bottom up or driven by the C-suite from the get-go. While eight months ago, companies adopted a number of different tools and employees may have used a personal ChatGPT license at work, that has shifted.
"Organizations are asking how AI needs to be built into workflows and which tools are actually best suited to gain the benefits in ROI," he said. "The businesses that are most successful have top-down vision and mandate, and then from the bottom-up is where you find the use cases. So there needs to be leadership around how and why you're adopting it, while from the bottom up, lots of use cases are being surfaced."
Herbert also said the level of education of AI among travel clients varies, with some companies already quite sophisticated and "pushing the boundaries" of what the technology can do.
"Others are still figuring out how to get started. The 'Eureka' moment for a lot of clients is when they see AI working with their own data and their own systems."